The use of survival time models with non-proportional hazard functions is introduced as a method of studying age of onset effects in family studies. Statistical methods for investigating non-proportionality of the hazard function are described, and the application of these methods is illustrated using data from a family study of depression to investigate transmission of age of onset between probands and their first degree relatives. The method is broadly applicable and useful for many chronic diseases where transmission of a disorder in families varies by age of onset of the disorder. It may also be used to investigate the effect of other proband factors such as sex on the age of onset of disease in relatives.